Pretty cool. I recently hatched nine baby walking sticks. The first one (and now the biggest) lost a leg somewhere along the way. I was bummed, but apparently that's quite common and she can grow a new one with each of her 3 skin sheddings. As Kramer so aptly pointed out, Mother Nature is quite the mad scientist!
BTW, watching walking sticks (and other insects in general) move around, cling and climb gives me a real appreciation for the biomechanics of those long, spindly legs. I guess daddy longlegs & skeeters really take the cake with even longer & skinnier legs.
I've run into this very problem with the floating ROVs my students are building. Sometimes, one of the thrusters will pick up some seaweed. Even though we've now got screens in front of the propellers, the seaweed gets stuck to the screen and the thruster performs less efficiently. In past trials, the ROV would then just spin around in circles, unable to adapt to the new situation.
What I've done to combat this is to keep track of every time the ROV is commanded to turn one way, but turns in the opposite direction, using the compass as a turn sensor. Each time this happens, the program increments or decrements a variable called turn_bias. This variable then gets added to the left thruster speed, and subtracted from the right thruster speed. After a short interval, it's able to move straight again and resume its desired trajectory.
Here's an image of one such trajectory:
The ROV picked up some seaweed on its right thruster almost immediately as it left the dock and began to curve to the right. After a few iterations back-and-forth adjusting the turn_bias variable (manifested as "scalloped" sub-trajectories), it was able to navigate normally again, with the seaweed still clinging to its right thruster when I brought it up on the dock.
Can this be uploaded to my wife to help her drive? Or heck, even to 75% of drivers. Though I prefer tasers in the chairs of those who drive slow in the fast lane.
It's interesting how they have it do much of the learning beforehand to allow it to try a few techniques to try and test at the time of failure.
Erco, are you into bugs? I just got into (exotic) roaches this year. I've got like a billion of them now. I use them for my research and teaching, but I have really grown to like them. They are way more fun than the bees my laboratory as traditionally used.
I would love to see an adaptive movement pattern change in a single long path instead of in repeats of many paths. Something more like that Phil posted (very cool!) is much more natural. Thinking about robots and bugs has me wondering if animals are rapidly learning to adapt to a loss of limb function, or if some of the injured movement patterns are a little more "instinctive." (Instinctive is in quotes because when you really get to look at behavior, it is not as useful of a concept as it first seemed.)
Not generally, I just like oddball things. Thus my Corvair, bassoon, Predicta, etc.. My teacher-neighbor gave us her class walking stick last summer, and our kids liked it. The ultimate low maintenance pet... some rose leaves every now and then! It died around Christmas, and the wife ordered some WS eggs from Turkey (!) which just recently hatched. I'd like to get some leaf insect (phyllium giganteum) eggs, but all the ones I found so far are in the UK and they won't ship here: http://www.ebay.co.uk/itm/Phyllium-mabantai-Leaf-Insect-Eggs-Ova-/131515231380
Comments
BTW, watching walking sticks (and other insects in general) move around, cling and climb gives me a real appreciation for the biomechanics of those long, spindly legs. I guess daddy longlegs & skeeters really take the cake with even longer & skinnier legs.
@erco - so are you pal!
What I've done to combat this is to keep track of every time the ROV is commanded to turn one way, but turns in the opposite direction, using the compass as a turn sensor. Each time this happens, the program increments or decrements a variable called turn_bias. This variable then gets added to the left thruster speed, and subtracted from the right thruster speed. After a short interval, it's able to move straight again and resume its desired trajectory.
Here's an image of one such trajectory:
The ROV picked up some seaweed on its right thruster almost immediately as it left the dock and began to curve to the right. After a few iterations back-and-forth adjusting the turn_bias variable (manifested as "scalloped" sub-trajectories), it was able to navigate normally again, with the seaweed still clinging to its right thruster when I brought it up on the dock.
-Phil
It's interesting how they have it do much of the learning beforehand to allow it to try a few techniques to try and test at the time of failure.
Publication - http://www.isir.upmc.fr/files/2015ACLI3468.pdf
I would love to see an adaptive movement pattern change in a single long path instead of in repeats of many paths. Something more like that Phil posted (very cool!) is much more natural. Thinking about robots and bugs has me wondering if animals are rapidly learning to adapt to a loss of limb function, or if some of the injured movement patterns are a little more "instinctive." (Instinctive is in quotes because when you really get to look at behavior, it is not as useful of a concept as it first seemed.)
Not generally, I just like oddball things. Thus my Corvair, bassoon, Predicta, etc.. My teacher-neighbor gave us her class walking stick last summer, and our kids liked it. The ultimate low maintenance pet... some rose leaves every now and then! It died around Christmas, and the wife ordered some WS eggs from Turkey (!) which just recently hatched. I'd like to get some leaf insect (phyllium giganteum) eggs, but all the ones I found so far are in the UK and they won't ship here: http://www.ebay.co.uk/itm/Phyllium-mabantai-Leaf-Insect-Eggs-Ova-/131515231380
@erco - amazing camouflage - and I come from the land of camouflage!
Check out the pink orchid mantis at http://www.keepinginsects.com/praying-mantis/general !